Breakthrough Ideas (continued)
India and All the Rest.
India.
India, with a population of 1.1 billion people, is Qualcomm’s fourth Asia card, the last Asian Ace in the hole. Indian bureaucracy is daunting, scaring away or wearing out many of the multinationals that made entries in the early 1990’s. Qualcomm’s strategy was to form an alliance with Reliance, India’s largest private firm. Reliance owns one of the world’s largest oil refineries, is one of the world’s top makers of polyester, and is one of the most profitable of Indian companies. Now Reliance is using its hard-charging business tactics and political clout to shake up the newly deregulated telecommunications industry. Although India has the second-largest population in the world, there are only 38 million fixed-line phones and 7.3 million cellular phones, arrayed in 42 networks across 21 telecom circles. This meant that wire line penetration is less the 4%, and mobile penetration, less than 1%. But, all these mobile phones were GSM.
Reliance saw the opportunity, and Qualcomm’s QSI was prepared to invest $200 million dollars in Reliance to facilitate its $5 billion dollar effort to build a telecom business that includes basic, mobile, long-distance, and broadband services, all using as needed Reliance’s new fiber optic trunking system linking 115 Indian cities. Reliance’s mobile service, which is categorized by bureaucrats as “limited mobility,” will use Qualcomm’s CDMA wireless local loop (WLL) technology. WLL permits landline operators to extend their networks. CDMA can provide high-density capacity at less than 1 cent per minute. Perhaps, up to 15-20 times AMPS cellular capacity or 4-5 times more than GSM. Third-generation fixed CDMA has a theoretical advantage of about 6 times mobile, but the current equipment and standards do not fully exploit the theoretical possibilities.
Reliance will install CDMA WLL in 19 circles in large cities, covering 90% of pops within its 25 km radius. Limited mobility means that handsets can be used only within the circle and cannot roam across circles.
WLL increases voice capacity because fixed, in particular, and limited mobility, to a lesser degree, not only permits the usual CDMA advantage of universal frequency reuse, but also permits very accurate tracking of power control that reduces overall transmission power, which, in turn, increases capacity. The near-far problem becomes more tractable; synchronization and orthogonality are easier to provide when people are not continually on the move. Also, each subscriber requires less RF power to achieve quality communication, meaning more subs per CDMA channel. There is less need for redundancy in base stations if congregating in crowded cells is not expected. All of these considerations reduce the number of cell sites substantially, lowering infrastructure costs. WLL simplifies RF planning and implementations, reducing build-out costs further. The CDMA interface still benefits from all its unique features such as variable or selectable mode vocoders, 1X’s doubling of network capacity, multipath immunity, soft handoff within each circle, and frequency, space, and time diversity that will continue to increase capacity with future upgrades. Perhaps, a WLL local loop might cost 50% less than full mobile for the same number of customers, with limited mobility gaining some fraction of those reduced costs.
Of course, the MSM6000 entry phone’s ZIF architecture, which reduced board size by 50% and BOM by 30-50%, permits the introduction of a sub US$50-70 handsets, which can more than compete with GSM handset prices because they will be rented at a lower rate. ZIF offset GSM’s advantage in economies of scale by a radical simplification of traditional heterodyne architecture that immediately made Qualcomm a strong competitor in low-end markets like India and Latin America.
As a new service provider, Reliance can quickly deploy its non-traditional WLL solution, with a soft roll out in September of this year and a rapid roll out next year. In fact, according to Thornley at SBB in September, “I mean in the millions of subscriber quite quickly; so it will become an important market for us in ’03.” Reliance expects to sign up 35 million subscribers by 2006.
In January, KTF, which won a $10 million dollar contract, reported that it will design the company’s RAN. In July, LG Electronics announced a $100 million handset contract from Alliance for 3 models of CDMA handsets. Alliance has been in talks with LGE, Samsung, and Nokia about handsets, which it will import or manufacture in India. Lucent, with a contract for 5 million CDMA lines, along with Nortel and Samsung were expected to win infrastructure contracts. Reliance said in a September interview that the networks were 60-70% complete.
The Telecom Regulatory Authority of India (TRAI) allows fixed-line operators, such as Reliance Infocomm, to offer mobile-phone services within city limits at rates equal to fixed-line local calls. Thus, “limited mobility” creates a financial bonanza for Alliance and other CDMA “limited mobility” providers. The largest player in India is government-run BSNL, with a 35 million fixed-line network. BSNL plans to roll out cellular services in more than 800 cities over the next year. Qualcomm reports that BSNL has issued a tender for 1.2 million-subscriber CDMA in 80 cities. The Indian conglomerate, Tata Teleservices plans a rollout of CDMA limited mobility services in the capitol, Delhi. TTSL’s operations include six circles: Andhra, Karnaataka, Tamil Nadu, Gujarat, Delhi, and Maharastra, including Mumbai. MTNL is a smaller carrier with 50,000 limited mobility CDMA lines in Mumbai and Delhi that are saturated, but it is hesitant to move forward until the regulator climate is clearer. These new carriers favor CDMA for cost effectiveness, capacity, and voice quality. But, they also appear to have a lock on the government that benefits them at the expense of GSM mobile operators.
In July, TRAI cut the basic fee over 50% for fixed-line carriers. Positioned as a poor man’s service, limited mobility is attracting a lot of interest. In a country with the lowest mobile tariffs in the world already, the rock bottom limited mobility prices are winning the support of policy makers and legislators. According to a July Reuters story from Mumbai [Bombay, the financial center] (emphasis added): Message 17766335
“Among the carrots being dangled to potential WiLL mobile subscribers are free incoming calls, tariffs a shade less than one cent per minute for outgoing calls while the monthly rental is pegged at a peak of $5. GSM cellular companies, in comparison, charge a minimum of one cent per minute for incoming calls and a minimum of 3.5 cents a minute for outgoing calls, while peak rentals touch $8.
In a price-sensitive market like India, that difference is enough to decide a winner and loser.
If the fixed-line companies are permitted to provide WiLL mobile service, they stand to make a killing. According to the revenue sharing agreement, if a call originates from a GSM network to a fixed-line phone, the former gets only 5 percent of each call, besides the airtime. The rest has to be passed on to the fixed-line company. Since WiLL mobile is licensed as a fixed-line company, but is offering mobile services, it will continue to get the lion’s share of the revenue. That effectively turns WiLL services into a mobile service, with the revenue benefits of fixed-line companies.”
Needless to say GSM carriers are furious and brought a lawsuit to the Supreme Court, where it was postponed in July and rescheduled to begin on September 18. Limited mobility companies continued their build-outs. Journalists report that experts expect limited mobility to prevail, as does Qualcomm. On September 2, Communications minister Pramod Mahajan said the DoT plans to reduce the license fee for basic and mobile service providers this year. Since telecom operators were shifted from fixed license fees to revenue sharing in 1999, prices decreased over 70%, among the lowest in the world. But, operating costs are among the highest in the world due to license fees and taxes. Mahjan said, “In the next 7-8 years, there will be a demand for between 10 and 15 crore [100 to 150 million] phones. This demand can be met only if there is a favourable licence fee regime and the price of services is low.” A cynic might say, “In India, it looks like rock bottom prices, but very favorable terms, are available to the highest bidder.”
Let’s imagine what Reliance’s strategy might be. Reliance segments the limited mobility market into groups based on income and life styles, say, Bare Bones, Young Hip, Middle Class, and Business Users. It provides either remaindered cdmaOne or new MSM6000 phones to the most price conscious segment, giving them voice and SMS services. Incidentally SMS service should be interoperable across CDMA and GSM networks. We assume the two middle groups are less price sensitive but differentiated by coolness versus pragmatic usefulness of applications. For the Young HIP and some middle-class, the MSM6500 will appeal for it BREW and Internet Launchpad features, including position location. Using any of these wireless applications increases the sophistication and comfort level of the subscribers. Over time, these groups and business users are potential wireless Internet customers. Business users are more likely to have some Internet experience and familiarity with computer applications. Here, mid- and high-end phones are likely to be either initial or replacement phones. Alliance can identify and help these subscriber become broadband users on 1X and 1xEV-DO limited mobility networks over a period of years. With their fiber optic network in place, they can build broadband demand from handset or PDA limited mobility users as well as from business users with PCs. PMCIA cards for the limited mobility network laptops or in home PCs are a natural. BREW makes this whole development process much easier. There are many programmers in Bangalore for writing or personalizing BREW applications to the Indian market. Just extending fiber optics to base stations or controllers reduces backhaul costs sharply.
My point is limited mobility users in India, who have never used the Internet or a PC, can become comfortable by using available wireless applications. Using the handset may be their first way or only means of becoming Internet users. High data rate wireless connections, friendly and useful devices, and applications tailored to customers make this possible. Entrepreneurs and users will join together to use these devices and networks to create new opportunities for global communication and accessing needed knowledge. Carriers will variously combine mobile wireless applications and broadband wireless Internet applications with other broadband possibilities and network, like Wi-Fi or other candidates as they emerge from innovative ideas. The cluster of related technologies grows, but mobile connections, always on, always with you, are always with us.
If India does not intend to eliminate GSM competitors, it must rebalance the scales, however. Of course, any GSM network failures might open the door for acquisitions and GSM1x installations if the economics permit.
Latin America.
CDMA has 14 million subscribers in South and Central America. CDMA2000. On September 9, the CDG reported that SMARTCOM PCS in Chile became the latest service provider to deploy 1X in Latin America. Nortel provided the $22 million in infrastructure. Three other Latin American operators had already launched 3G services: Centennial de Puerto Rico (coverage of Puerto Rico, U.S.Virgin Islands, Dominican Republic, and Jamaica; on September 13, Bermuda’s CellularONE announced in entered an agreement to launch a CDMA2000 1X network in March 2003.), in April; Telefonica Celular of Brazil in April, and Telesp Celular, the largest mobile operator in Brazil, in December 2001. In July, Vesper of Brazil announced a planed 1xEV-DO network. Qualcomm holds a 16% interest.
Of the nine conversions from TDMA to CDMA, four are in Latin America. Movilnet is deploying a CDMA2000 1X overlay in Venezuela. In July, BellSouth Ecuador selected Nortel to deploy Ecuador’s first CDMA2000 network as an overlay of TDMA. In August, Ericsson signed an agreement with BellSouth Panama for another 1X overlay. However, the first 3G 1X deployment in Central America will be in Guatemala, with Nortel providing the equipment. Both BellSouth International and Verizon International plan to convert additional TDMA networks to 1X.
Telesp Celular launched a BREW user trial in June. BREW MOUs have been signed with BellSouth in Guatemala, Telcel-BellSouth in Venezuela, and Vesper in Brazil. Although there are signs of continuing activity, in September, Dr. Jacob’s said, “I think the slow market will continue to be Latin America – very difficult economics there, and elections are coming in Brazil, so everybody is holding back.” Obviously, low cost phones are essential in this market, meaning the MSM5105 or the cheaper MSM6000, both 1X Rel 0, voice only phones, with 2-way SMS.
Of course, the standards war has broken out in Latin America, with 3GAmericas maintaining that GSM will be the most popular technology in Latin America in 2006, when it will reach 48% of the 178 million subscriber-market according to Pyramid [scheme?] Research. Qualcomm insists that this will not happen given that CDMA has 4 times the number of subscribers than GSM now, 22% to 5 %, and in the last year gained 3 times more than GSM did, 6 million to 2 million.
Asia/Pacific.
On July 22 2002, Telecom New Zealand launched the first CDMA1000 1X network in the region. In 1999, the grip that GSM held on the market was broken when Telstra of Australia and then Telecom of New Zealand decided to convert its TDMA network, which was an overlay of its 1987 AMPS network, by deploying cdmaOne as an overlay. For TNZ, it was the only way it could mount a competitive challenge to Vodaphone NZ, the country’s only GSM carrier. For Telstra, despite its own commitment to GSM, building a CDMA network was necessary to meet its obligation to the Australian government to provide coverage in the vast outback. Results exceeded their expectations. For TNZ, its AMPS network had taken six years to attract 100T subscribers, with CDMA that number was reached in six months, with half upgrades and half new customers.
The TNZ 1X network covers 97% of the population of 4 million, with over 1m in Auckland. As TNZ Mobile General Manager Lorraine Witten put it, “The flexibility of the CDMA network provides us with an easy and cost-effective way to deliver benefits such as longer standby times, peak data rates of up to 153 kbps and affordable upgrade paths to our customers.” In trials, the average throughput was 50 to 80 kbps, competitive with dial-up modems. Twenty devices, including two handsets, the Sony Ericsson t60c and the Kyocera 2235, complement the GTRAN Dot Surfer card in use for some time. (The GTRAN 1X card for laptops, PDAs and other mobile devices is in use in 8 countries: Korea, Venezuela, Peru, Brazil, China, U.S., New Zealand, and Puerto Rico) According to Corporate and Business Head, Warwick Beban, Telecom’s launch of Mobile Jet Stream services means, “people can, for the first time, really work remotely – places other than home or work, and use high-speed services such as mobile office applications, accessing email and Internet. In a week or two they’ll also be able to use their phone to take and send photos. Mobile Jet Stream is a step ahead of Vodaphone’s GPRS offering, which should be compared more accurately with our existing CDMA service. The user experience on Mobile Jet Stream is akin to that of Telecom’s fixed line JetStream Starter Service, which achieves speeds up to 128kbps.”
In April, Telstra, the biggest wireless operator in Australia, is extending its CDMA coverage to 132 towns with over 500 residents; Telstra also received an US$11 million dollar government contract to extend its CDMA coverage across 2,600 km of Australian highways. Telstra’s cdmaOne network covers 1.3 mm square km, the second largest in the world. A recent survey of customers found that 81% would recommend CDMA to their friends. Teslstra has been bullish on CDMA 2000 trials with Nortel, which can upgrade to 1X within six months for a cost of around A$105 million (exchange rate 55 cents to US$1). “Nortel estimated that the cost to deploy such a network across Australia would be similar to the US cost of US$3 per head of population.” Lucent did a study for Telstra advocating CDMA 1X in a 450 MHz band to take maximum advantage of that frequency’s long-range coverage, which requires far fewer cell sites than higher frequency bands, for the outback. On September 11, the Wall Street Journal reported this quote from David Thodey, managing director of Telstra Mobile, “It’s a real possibility for us. Korea and KDDI have show there’s a really strong opportunity there.”
Upcoming 1X launches in Asia include, Telekom Malaysia in the 450MHz band, Ratelindo in Indonesia, and Hutchinson CAT in Thailand.
North America.
In North America, there have been 7 launches of CDMA 2000 1X: Monet Mobile (USA), 8/4/01; Leap Wireless (Cricket, USA), 12/10/01;Verizon (USA), 1/29/02; Metro PCS (USA), 2/1/02; Bell Mobility (Canada), 2/12/02; Telus Mobility (Canada), 6/3/02; and, Sprint PCS (USA), 8/12/02. As of March ’02, Verizon, with 23.7 million CDMA subscribers, was the world’s #1, Sprint, with 15.8 million CDMA subscribers, was the world’s #3; Alltell, with 4.7 million CDMA subscribers, was the world’s # 7; Bell Mobility, with 1.8 million CDMA subscribers, was the world’s #10; Telus, with 1.5 million CDMA subscribers, was the world’s #12; Leap Wireless, with 1.4 million CDMA subscribers, was the world’s #14; Qwest, with 1.1 million CDMA subscribers, was the world’s #15; and other carriers had les than 1M. Thus, the North America, with 50 million subscribers, has the largest CDMA market, followed by Korea, with 30.4 million, and Japan, 10.8 million. Gartner Dataquest estimates there will be 65.6 million subscribers in North America by the end of CY’02.
Qualcomm expects 80-85 million MSMs or handset sales for FY’02 ending in September, with about 70% replacement sales, driving the N.A. Market into the 40-42 million range, and Asia, mainly Korea and Japan, accounting for 30 million. Some analysts have remained skeptical in spite of Qualcomm claiming that 95% of that estimate was in hand in July. Thus, the present quarter will give you a good sense of Qualcomm’s accuracy, which is bounded by 15 (for estimated forward guidance) to 19 million MSMs, a record, which they, in fact, expect.
Verizon, Lucent, and Qualcomm have worked closely together in its staged rollout of 1X. In August, Lucent introduced two new base station systems that improve on existing technology, with a smaller footprint, for 1X and Verizon’s 1xEV-DO trials. Verizon, who currently claims 29.4 million customers, up 5.7 million from the March ’02 figure, will be offering 5 color screen phones, 4 with gpsOne in October for the Xmas push. From early trials, they see data and voice ARPU go up with color screens and BREW. They believe high-end handsets will not have to be subsidized as much, lowering CAPC. Evidence of this effect is also present in Korea, both for ARPUs and CAPC, where the government ban on handset subsidies did not impact sales after the first month. Comparisons over the next year between Verizon, with BREW, and Sprint, without BREW, should reveal the impact of BREW on voice and data ARPUs, assuming some offset for Sprint’s strong emphasis on Enterprise markets. The most recent update on BREW is 300 applications on 19 handsets, with 50 more devices in development, and 1mm BREW handsets downloading applications. Qualcomm expects position location services to drive the market. Also, Verizon’s success with its DO trials is crucial for future rollouts.
Because Sprint was selling 1X devices before its network rollout, it had 1.3mm 1X devices already in place. Sprint is offering packages of bundled voice and data minutes, include introductory prices--$90-per-month for 2 MB of data and 2,000 minutes of voice airtime, with one- or two-year contract; for data only package offers unlimited for $50 per month for the first 90 days, and $100 per month thereafter with a one-year contract. Although Sprint offers more aggressive pricing, these prices are still not as low as may be necessary to spike strong demand for data services. Its emphasis is on becoming the dominant wireless data provider to the enterprise, providing peak performance of 144 kbps and expected average data speeds of 50-70 kbps. PCS Vision’s initial 3G handsets included Handspring’s Treo, a Toshiba Pocket PC platform handhelds, plus digital cameras from Sanyo, Samsung, Hitachi, and LG.
Sprint President Charles Levine described the 1X launch as the biggest event since the initial launch of CDMA in the mid-1990s, “giving us competitive advantage that we haven’t had in awhile.” The new network results in a crucial improvement in voice services by doubling its capacity, given that enough customers buy 1X-enabled phones. “We took a $10 billion plus network and for $1.4 billion we doubled its capacity.” The Sprint network is covered by 18,325 BTSs. The company plans to use $3 billion in 2003 cap ex to bolster its data network speed. Levine hopes to see more than 20% of Sprint’s customers using data services by the end of 2003.
Sprint has no plans to deploy 1xEV-DO, saying that it will wait until 1xEV-DV becomes available. This is not a standard that Qualcomm endorses because it believes spectral efficiency is best when voice and data are separated into distinct 1.25 MHz channels, each optimized for either voice or data delivery. Sprint’s most pressing need may be to hold down capital expenditures. Sometimes Sprint spokespersons say they will not be locked into unfilled data channels, as if do not understand that these separate channels can be allocated between voice and data minute by minute. Once Sprint establishes that there is a viable data market, I believe that Qualcomm expects them to add DO and BREW because they are complete standardized solutions that can drive the data market.
Using minutes purchased from Sprint PCS, Virgin Mobile USA became the first cdmaOne mobile virtual network operator (MVNO), offering prepaid service, topped up by credit cards, for a targeted youth market of American teens that has proved to be so successful in Europe. With a fixed rate of 25 cents for the first 10 minutes, with ten cents a minute after that, Branson expect better than 350,000 sales by the end of the year. This plan could produce some happy teens and unhappy parents when the credit card bills come rolling in.
Both Leap Wireless and MetroPCS, at $35-39 per month, offer prepaid plans low-end customer offering all the minutes you can eat within a local market. At the moment the macroeconomic environment is difficult, given the costs of building out its networks, but Leap’s new networks, available in 40 markets in 20 states, in second-tier U.S. cities are becoming EBITDA positive within 13 months. The approach uses cost savings from cheap voice capacity, prepayment, and careful market surveying to rollout cookie cutter duplicates in new local networks. The strategy is to compete with and disrupt local wire line companies. Leap is leading the wireless industry in driving the trend of landline displacement, with 26% of its customers having cut the cord. About 2.6% of U.S. consumer “cut the wire” in 2002. IDC says, 13.8% are “interested” or “very interested” in doing so. In comparison to wired, wireless voice minutes are rising steadily, from 7% in ’99, 10% in ’00, 15% in ’01, 20% this year, and, perhaps, 28% in ’05. In other countries, the switchover has occurred. For example, Hungary has 5.86 million cell phone users versus 3.72 million wireline. In China wireless users will exceed wireline users in the first half of ’03. The disruption of wireline by wireless worldwide is generally accepted as a nearly inevitable long-term market trend.
On September 11, Spain’s Telefonica Moviles finalized the merger with Pegaso, Mexico’s #2 player in an expanding market. This means Qualcomm receives debt repayment with 60 days. |
